Safe Withdrawal Rate at 3% Inflation: Is 4% Still Safe?

No, the 4% rule isn’t a law, especially with 3% inflation

No, the 4% rule isn’t a law, especially with 3% inflation. The most persistent myth I still hear, on golf carts, at barbecues, and yep, even in boardrooms, is that you can set a 4% withdrawal and coast for 30 years. That’s not how markets, or spending, work. The original “4% rule” came from William Bengen’s 1994 paper using U.S. data back to the 1920s, assuming a 30-year retirement, annual inflation adjustments, and a roughly 50-75% equity allocation. It was a result under those exact inputs. Useful starting point. Not a guarantee.

And inflation matters. A lot. With inflation running around ~3% as a planning input this year (BLS 12-month CPI has hovered near 3% for much of 2025), a 4% rule means your real spending stays flat only if you actually increase your withdrawals by that 3% each year. Quick napkin math: start with $1,000,000, take $40,000 in year one, and with 3% inflation you’re at $41,200 in year two; by year ten you’re near $52,000. Simple? Sure. But the market doesn’t care about your spreadsheet compounding cleanly…

Sequence-of-returns risk is the real boss. Average returns lie; the path is what drains accounts. Two quick reminders from history: the S&P 500 fell about 49% from 2000-2002 and ~57% from 2007-2009, while a classic 60/40 portfolio dropped roughly ~20% across 2000-2002 and about ~22% in 2008 alone (stocks ~-37%, core bonds ~+5% that year). If those hits arrive early in retirement while you’re pulling cash, the math gets ugly. Same average return over 30 years, radically different outcomes depending on the first 5.

And fees and taxes? They quietly shave the safe rate. A 1% all-in fee could reduce sustainable withdrawals by roughly 0.5-1.0 percentage point, based on multiple retirement research runs since 2010 (Wade Pfau, Michael Kitces, and follow-on Bengen work). Taxes turn “gross” 4% into something less after the IRS gets a turn. Not trying to be dramatic, just realistic.

Here’s what we’re going to do with 3% inflation as our anchor. We’ll build a plan that survives ugly markets, not just pretty charts. That means embracing flexibility, variable withdrawals, guardrails, cash buffers, and aligning allocation with your tax picture and fee drag. I’ll circle back on the inflation point because it’s the fulcrum: with 3% inflation, rigid 4% can be fine in some stretches and way too aggressive right after a bear market. Or too conservative when yields are decent and your spending is flexible. Yes, both things can be true. It’s gray.

What you’ll get from this section:

• Why 4% is a starting line, not a finish line. Bengen (1994) was conditional, not universal.
• How 3% inflation changes the math on real spending growth and your paycheck from the portfolio.
• Why flexibility beats rigidity in the real world, guardrails > autopilot.
• How sequence risk can dominate averages, with real market context from the 2000s and 2008.

One last thing, because I over-explain this and then regret not being briefer. A 4% initial draw adjusted for CPI is a rule about maintaining purchasing power, not about maximizing the account. If markets are choppy, you may trim raises or pause inflation adjustments. If yields are stronger, 10-year Treasuries have spent much of 2025 around the mid-4% area, different story. Okay, that’s the setup. Now we’ll make it practical…

What 3% inflation actually means for your retirement paycheck

Inflation isn’t a vibe; it’s the raise your spending demands every year whether markets cooperate or not. At 3%, the Rule of 72 says prices roughly double in about 24 years (72 ÷ 3 ≈ 24). That’s not theory, historically, U.S. CPI has averaged about 3% over the very long run; the Ibbotson/SBBI long-run series (1926-2023) pegs it right around that 3% mark. Which is why 3% is a reasonable base case for planning, not a bearish call or a guess about next month’s CPI print.

Here’s the translation to your paycheck: if you want to keep your real lifestyle stable, a $80,000 first-year withdrawal needs to step up to $82,400 next year, then about $84,872 the year after, and so on. That compounding is sneaky, after 10 years at 3% inflation, a flat $80,000 payment only buys about $59,500 of today’s goods and services because 1 / 1.03^10 ≈ 0.744. That’s a ~26% loss of purchasing power in a decade. Stretch it to 20 years and you’ve lost ~45% (1 / 1.03^20 ≈ 0.553). I’ve seen very solid balance sheets get pinched not because returns were terrible, but because people stuck with level-dollar withdrawals and quietly fell behind on real spending.

And yes, I know this can get a bit mathy, sorry, I said “real” spending and that’s econ-speak for inflation-adjusted. Simple version: if you don’t give yourself a raise each year that at least matches inflation, you are giving yourself a pay cut in what your money actually buys. The catch is markets aren’t a straight line, and sometimes the best move is to accept a smaller raise or even a temporary pause to protect the plan’s durability.

That’s where dynamic raises earn their keep. There are practical rules, guardrails, that say: take your baseline inflation raise most years, but if your withdrawal rate (spending ÷ portfolio) drifts too high after a rough market, pause or cap the raise. If your withdrawal rate drops back into the safe zone after a rebound, resume normal inflation adjustments. It’s not fancy; it’s just giving the portfolio breathing room in bad sequences.

Context matters. Yields this year have been more helpful than they were in the zero-rate era, 10-year Treasuries have hovered in the mid-4% area for much of 2025, which supports the arithmetic of inflation-adjusted withdrawals a lot better than 2020-2021 did. Still, a 3% spending escalator is a real hurdle for the portfolio to clear, especially early in retirement when sequence risk is loud. A bad first five years plus automatic 3% raises can chew through capital faster than you expect.

So my take, admittedly one guy’s opinion after two decades of sitting with families and spreadsheets, is:

• Anchor to 3% as your default inflation adjustment because history supports it (SBBI 1926-2023 ≈ 3%).
• Run the numbers on purchasing power: a flat paycheck loses ~26% in 10 years and ~45% in 20 at 3% inflation.
• Use dynamic raises to stabilize the plan in tough markets: pause or trim raises when withdrawal rate bands are breached, then catch up later if/when markets cooperate.
• Pair raises with yield reality: when bond income improves (as in 2025), you’ve got more room; when yields compress, be tighter.

I’ve seen more plans struggle from ignoring compounding inflation than from a single bad year of returns. Slow erosion is quieter, and that’s why it’s dangerous.

It’s messy, because life is messy, healthcare spikes, travel ebbs, roofs leak, but the principle holds: plan for a 3% raise as the default to keep your lifestyle intact, then be flexible about when and how you take it. That blend of discipline plus wiggle room is what keeps the paycheck real without breaking the portfolio.

So what’s a safe starting rate at 3% inflation?

Short answer: it depends. On horizon, mix, fees, taxes, and whether you can nudge spending when markets throw a tantrum. Historically, Bill Bengen (1994) found roughly 4% worked over 30 years in U.S. data when you fully inflation-adjust every year and held a diversified stock/bond mix. That’s the old anchor. Today, with 3% as your planning inflation and a 50-75% equity mix, a realistic starting range for a 30-year horizon is about 3.6%-4.7% before fees and taxes, lower if you want near-certainty, higher if you’ll trim raises or skip a raise in rough patches.

Two quick market checks for context in 2025: U.S. 10-year Treasuries are hanging in the mid-4% range as we head through Q4 2025, and long TIPS real yields have hovered around ~2% this year (the 30-year TIPS real yield was near 2.2%-2.4% for much of mid-2025). Those real yields raise the floor versus the 2010s, when real yields were often negative, which helps the case for the upper half of that 3.6%-4.7% band if your equity mix is balanced and you’re not fee-heavy.

But, and I say this as someone who’s watched too many “rules” break on contact with reality, horizon matters. If you’re planning for 40 years, I’d trim roughly 0.3-0.5 percentage points off that 30-year range. So think something like the low-3s to low-4s before fees/taxes, with the same caveat: more flexibility, more room.

Fees and taxes: every ongoing drag tightens the belt. All-in costs of 0.75%-1.0% a year (advisory + fund + trading + taxes from turnover) can knock your sustainable starting rate down by ~0.5-1.0 percentage points in long-horizon simulations. I know, that sounds linear. It isn’t perfectly linear, but it’s close enough for planning that you should budget for it. And if you’re in a high tax bracket drawing from taxable accounts first, be conservative on the starting number.

Dynamic rules vs rigid CPI raises: classic “4% then full CPI every year” is simple, but it’s the least forgiving. Guardrail systems (e.g., bands on the withdrawal rate with rules to skip or trim raises) historically supported higher starts in U.S. data. Guyton-Klinger (2006) showed that with decision rules, starting rates around ~4.5%-5% were often supportable over 30-year horizons, before fees/taxes, with spending cuts when bands were breached. Reality check: those results depend on market history and willingness to accept variability. If you’ll actually pause raises when your withdrawal rate drifts above, say, 5.5%-6% of assets, you can usually start a bit higher than a rigid CPI-only plan without losing sleep.

If I synthesize it for 2025: for a 30-year plan, 50-75% equities, 3% inflation planning, reasonable costs, I’d frame it like this, 3.6% if you want near-bulletproof, ~4.0%-4.5% for most flexible planners, up to ~4.7% if you’ll actually use guardrails. For 40 years, shave off 0.3-0.5 points. And yes, I know that’s messy. Retirement is messy. The trick is matching the starting rate to your flexibility, not to the cleanest-looking spreadsheet.

Small personal note: the clients who slept best last year were the ones who skipped a raise in 2022-2023 and then took a partial catch-up earlier this year when yields improved. Imperfect, but effective.

Sequence risk still bites: lessons from 2000-2002, 2008, and 2022

The problem isn’t average returns; it’s the bad early years. I know that sounds obvious, but it keeps showing up in client files. If you start withdrawals and your first two or three years are ugly, a fixed 4% + CPI raise can turn fragile fast. The math is annoyingly simple: you’re pulling cash out while prices are down, which leaves fewer shares to rebound. Then your next raise is larger (because inflation flares when it’s least convenient), so you yank a little more from a smaller base. Rinse, repeat, and suddenly the long-term average return you were promised feels like a fairy tale.

We’ve had real-world stress tests that prove the point:

• 2000-2002: The S&P 500 Total Return fell three straight years: about -9.1% (2000), -11.9% (2001), and -22.1% (2002). That’s roughly -37% cumulative before the recovery even started. If you began retirement in 2000 and stuck to 4% plus raises, you were selling into a slide for 36 months.
• 2008: The S&P 500 Total Return was about -37.0%. Bonds helped that time, the Bloomberg U.S. Aggregate Bond Index returned roughly +5.2% in 2008, which gave rebalancers some dry powder.
• 2022: Brutal because both sides fell. The S&P 500 Total Return was about -18.1%, and the Bloomberg U.S. Aggregate lost roughly -13.0%, the worst bond year in decades. CPI inflation averaged about 8.0% in 2022, so anyone giving themselves a full CPI raise pulled even more from a shrinking portfolio.

Over-explaining a simple thing here: imagine $1,000,000, a 4% first-year withdrawal ($40,000), and then a -20% market year. You end near ~$760,000 after spending, not counting fees and taxes. If inflation pops 8% the next year, your withdrawal goes to ~$43,200, now that’s ~5.7% of the new balance. Repeat that twice and the damage compounds, even if the long-run average is fine on a spreadsheet.

So I care less about the exact average return and more about your Plan B. Small trims go a long way. I’ve watched clients who flex 5-10% on spending sail through the same markets that forced rigid plans to the brink. And it’s not heroic stuff, skip the full raise, delay the car by a year, pause the big trip. After 2022, several retirees skipped raises in 2023, then took a partial catch-up earlier this year when yields were north of 5% on T‑bills. That tiny bit of give protected their principal while income from cash did some heavy lifting.

Tools that help when the sequence is bad:

• Guardrail trims: Pre-commit to pause or cut 5-10% if your withdrawal rate crosses a threshold (say 5.5%-6% of assets). It’s a nudge, not a diet.
• Cash buffer: Hold 6-24 months of spending in cash or short T‑bills. As of Q4 2025, 3‑month T‑bills are still around the mid‑5% area, so the opportunity cost isn’t painful. That buffer lets you avoid selling stocks into a hole.
• Rebalancing: After drops like 2000-2002 or 2008, systematic rebalancing buys what’s down. In 2008, bonds were positive, which made the buy even easier. In 2022, fewer safe havens worked, but rebalancing still kept risk on target.
• Skip-a-raise policy: If CPI is hot (like 2022’s ~8% average), take half the raise or none until portfolio returns catch up.

Is it messy? Yep. Markets don’t move in neat annual increments and your spending isn’t a straight line either. But those small, temporary cuts are often the difference between a plan that survives a nasty start and one that spends a decade digging out.

Personal note: the retirees who slept best last year were the ones who said, “we can wait.” A 5-10% trim for a year or two beats a 25% cut later. I’ve seen that movie enough times to know which version I’d rather watch.

2025 reality check: yields, fees, and taxes matter more than your spreadsheet

Here’s the good news this year: real yields are finally doing their job again. As of late October 2025, 10‑year TIPS are sitting around ~2.1%-2.3% real (check the Fed’s H.15 data), versus the 2010s when they spent long stretches near 0% or even negative. That shift sounds small, but it’s not. A 2% real anchor means your fixed income actually helps fund inflation‑adjusted withdrawals instead of just babysitting volatility. For retirees running a classic 60/40, that higher real carry gives you more cushion than you had five or eight years ago.

Cash is still competitive as we head through Q4. Three‑ to six‑month T‑bills have been hovering roughly ~4.8%-5.2% for much of Q3 and into October (again, H.15). But don’t over‑park. That 5% can feel addictive, I get it. The catch is reinvestment risk. If the Fed’s path pulls bills down later this year or into 2026 and you roll from 5.0% to, say, 3.5%, that’s a ~30% pay cut on your “safe” income stream. Laddering and extending duration gradually, locking some real yield with 5-10 year TIPS or intermediate IG corporates, reduces that whiplash.

One more reality: fees and taxes quietly shave 0.5%-1.5%+ off the rate you think you’re earning. And the safe withdrawal rate that matters is the net rate after those drags. Quick math to make it tangible:

• Example: $1,000,000 in short Treasuries at 5.0% = $50,000 interest. A 0.40% advisory/ETF fee is $4,000. At a 22% federal bracket on ordinary income, taxes are $11,000. Net = ~$35,000, or 3.5%. Add 5% state tax and you’re closer to ~3.0%-3.3%.
• If you swap some bills for 10‑year TIPS at ~2.2% real, your after‑inflation spending power is clearer: that’s ~2.2% before fees/taxes. Net might be ~1.5%-1.8% real depending on costs and brackets.

That’s why asset location and bracket management aren’t “nice to have” in 2025, they’re must‑do’s:

• Asset location: Put taxable‑interest stuff (T‑bills, IG bonds, TIPS funds) in IRAs/401(k)s when you can. Keep broad equity index funds with qualified dividends and long‑term gains in taxable to use 0%/15%/20% rates.
• NIIT watch: The 3.8% Net Investment Income Tax still kicks in at $200k single / $250k MFJ MAGI (ACA thresholds, not indexed). A poorly timed capital gain can turn a 15% rate into 18.8% for the slice above the line.
• IRMAA cliffs: Medicare Part B/D surcharges are cliffy. In 2024, the first IRMAA tier started at $103k single / $206k MFJ MAGI; 2025 thresholds are a bit higher, but it’s still a cliff. One extra dollar can add hundreds per month, two years later. Careful with Roth conversions and RMD bunching.
• Harvest smart: Tax‑loss harvest when it’s there; don’t force it. Use losses to offset gains and NIIT exposure. On the flip side, harvest gains in the 0% bracket years.

If you like rules of thumb: higher real yields than the 2010s support safer withdrawals, but your safe percent is the net percent. Fees, taxes, and reinvestment risk are the levers that shrink or expand that number. I’ve seen plenty of plans where trimming costs by 40 bps and shifting bonds into tax‑deferred bought as much spending room as an extra 5% in stocks. Not glamorous, just effective.

Personal note: I love spreadsheets too, but your 4.0% line looks different when the IRS and a 0.40% fee take their cut. Get the location right, secure some real yield, and let the cash bucket be a tool, not a trap.

Frequently Asked Questions

Q: Should I worry about 3% inflation breaking the 4% rule this year?

A: Short answer: yes, a little. The 4% rule assumed you raise withdrawals with inflation, so with ~3% CPI running this year, $40,000 becomes ~$41,200 next year on a $1M nest egg. The catch is markets aren’t smooth. If returns are weak early, those inflation bumps can bite. Keep the raise flexible, tie it to portfolio health, not just the CPI print.

Q: What’s the difference between the classic 4% rule and a dynamic withdrawal plan?

A: The 4% rule is a simple starting point from Bengen’s 1994 work: take 4% in year one, then raise that dollar amount by inflation each year, assuming roughly 50-75% in stocks and a 30-year horizon. Dynamic plans adjust. You set guardrails: spend more after strong markets, trim or pause inflation raises after weak ones. Approaches like Guyton‑Klinger or “floor-and-upside” (pension/Social Security/TIPS as floor, equities for growth) handle sequence risk better and can keep you solvent when the first 5 years are ugly. Fewer spreadsheets, more reality.

Q: Is it better to start retirement with a 60/40 or 70/30 portfolio if I’m aiming near a 4% withdrawal?

A: It depends on your stomach and cash needs. Historically, equity risk drives long‑run sustainability, but it also amplifies early‑retirement drawdowns. A 60/40 fell roughly ~20% across 2000-2002 and about ~22% in 2008; a higher‑equity mix would’ve felt worse then. If you want close to 4% with inflation raises, many retirees blend: 55-65% equities, a 2-3 year cash/short‑bond bucket, and rebalance rules. Lower fees and smart tax sequencing can matter as much as tweaking 60/40 vs 70/30, no joke.

Q: How do I build a withdrawal plan that can survive a bad first five years of retirement?

A: Think of it as engineering around sequence risk, the nasty scenario where markets drop early while you’re pulling cash. Here’s a practical framework I use with clients (and yeah, I’ve seen this movie in 2000-2002 and 2008):

1. Right‑size the initial rate: If 3% inflation is your planning input this year, a 3.5-4.0% starting withdrawal is only reasonable if you’ll cut or pause raises after down years. Conservative households start 3.0-3.5% and give themselves raises when markets cooperate.

2. Build a cash buffer: Hold 2-3 years of withdrawals in cash/short Treasuries so a 2000-2002 (S&P ~‑49%) or 2008 (stocks ~‑37%, core bonds ~+5%) doesn’t force stock sales at fire‑sale prices. Spend the cash in bad years; refill it after gains.

3. Use guardrails: Adopt rules like Guyton‑Klinger, pause inflation increases after a negative year, cut spending 5-10% if the withdrawal rate (spend/portfolio) drifts above, say, 5.5-6.0%, and allow raises after strong years. Not fancy, just disciplined.

4. Lower friction: A 1% all‑in fee can reduce sustainable withdrawals by roughly 0.5-1.0 percentage point per research from Pfau/Kitces/Bengen follow‑ups. Negotiate fees, prefer low‑cost funds, and rebalance with bands to avoid over‑trading.

5. Tax order matters: Usually spend taxable first (harvest gains up to the 0% or 15% bracket), then traditional IRAs, save Roth for last. In gap years, do partial Roth conversions. Use QCDs after age 70½ to satisfy RMDs tax‑efficiently. Taxes quietly turn “4% gross” into 3.x% net.

6. Add resilience: Consider a small TIPS ladder for near‑term spending, or a modest deferred income annuity as a longevity floor. Stress‑test with historical sequences and a Monte Carlo. If the first five years come in ugly, your rules should auto‑tighten spending. If they’re decent, you can give yourself that inflation raise without sweating every CPI release.

Do this, and the path, not just the average, works in your favor. And yes, keep the plan in writing; future‑you will thank present‑you.

@article{safe-withdrawal-rate-at-3-inflation-is-4-still-safe,
    title   = {Safe Withdrawal Rate at 3% Inflation: Is 4% Still Safe?},
    author  = {Beeri Sparks},
    year    = {2025},
    journal = {Bankpointe},
    url     = {https://bankpointe.com/articles/safe-withdrawal-rate-3-inflation/}
}